基于图像识别的地毯机运动轨迹规划研究

Trajectory planning of carpet machine based on image recognition

为了改变人工描绘地毯机运动轨迹的生产方式,提出了一种基于图像识别的轨迹规划算法,该算法包含颜色量化、颜色聚类、边缘检测、轮廓跟踪和区域填充5个部分。在CIELAB颜色空间对彩色图像进行量化和聚类,对得到的聚类图像进行边缘检测和基于Freeman链码的轮廓跟踪,并采用扫描线填充算法进行区域填充,最后根据轮廓点坐标和填充点坐标组成地毯机的运动轨迹。实验结果表明:基于图像识别的轨迹规划算法实现了图像的轨迹规划,验证了算法的正确性。

In an attempt to change the production mode of depicting motion trajectory of carpet machine manually,a trajectory planning algorithm based on image recognition is proposed,which includes five parts:color quantization,color clustering,edge detection,contour tracking and area filling.The color images were quantized and clustered in CIELAB color space.Edge detection and contour tracking based on Freeman chain code are carried out for the clustering images.The area is filled by scanning line filling algorithm.The motion trajectory of the carpet machine is composed by contour point coordinates and filling point coordinates.The simulation results show that the algorithm based on image recognition realizes automatic trajectory planning and verifies the feasibility of the algorithm.